D2 — follow-up to the sibling-merged D1

Sibling PR 81d08c42 (D1, merged earlier today) mirrored OGAR #110's auth
domain. OGAR #111 just merged on the same day, adding two more entries
to ogar_vocab::class_ids::ALL:

• product 0x0207 — product.template / product.product converge here
• accounting_account 0x0208 — account.account / account.account.template
  converge here

(Same convergence pattern as account.move ↔ sale.order → commercial_document.)

Without this sync, lance_graph_ogar::parity::COUNT_FUSE const-eval-panics
the moment a consumer bumps its OGAR pin to current main: mirror=43, ALL=45.
Same shape as the q2 #42 trigger that exposed D1's drift.

Background — D #596 closed in favor of this

When PR #596 (my original D, attempted to fix the OGAR #110 mirror) tried
to rebase onto main, it conflicted with the now-merged sibling D1
(81d08c42). #596 is closed; this PR (D2) is the new follow-up, scoped
to the next drift (OGAR #111) using the discipline named in
E-OGAR-AUTH-MIRROR-DRIFT.

Two coordinated fixes

1. lance-graph-contract: extend ogar_codebook::CODEBOOK with the 2 new
   commerce-arm entries. No new ConceptDomain variant — both stay Commerce
   (0x02XX). 6 lines added.

2. Workspace Cargo.lock: bump pinned OGAR crates (ogar-vocab,
   ogar-class-view, ogar-ontology, ogar-adapter-surrealql) from
   08a9c979 → ac4b4162 (current OGAR main, post-docs: add γ+φ encoding + ModelPipeline DTO + Jina eval to handover Complete next-session pipeline: 1. Temperature fix (nucleus + thinking styles) 2. γ+φ golden ratio HDR re-encoding (per-role, Gate γ=1.50) 3. StandardizedModelPipeline DTO (6 models) 4. Jina cross-model eval (truth anchor) 5. Maverick 128E streaming (800 GB → ~1.2 GB) https://claude.ai/code/session_01ChLvBfpJS8dQhHxRD4pYNp #111 merge). 4 refs.

Safe SHA swap verified — Cargo.toml deps byte-identical between the two
commits:

ogar-vocab               0
ogar-class-view          0
ogar-ontology            0
ogar-adapter-surrealql   0

Tests

• cargo test -p lance-graph-contract: 9 doctests pass (compile-time fuse passes)
• cargo test --lib in crates/lance-graph-ogar: 53/53 pass, including
  parity::tests::mirror_is_a_faithful_copy_of_ogar_codebook
• cargo build -p lance-graph-ogar: clean

Discipline working as designed

This PR demonstrates the E-OGAR-AUTH-MIRROR-DRIFT epiphany (in EPIPHANIES.md
since sibling D1) holding: OGAR #111 merged this morning, the mirror catches
up this afternoon (instead of three days later when q2's lock bumps). Same
PR shape as D1; same minimal scope.

Source-of-finding

• feat(vocab+class-view): mint PRODUCT (0x0207) + ACCOUNTING_ACCOUNT (0x0208); extend ODOO_ALIASES OGAR#111 (introduced PRODUCT + ACCOUNTING_ACCOUNT; closes C of A+B+C)
• AdaWorldAPI/odoo-rs#14 (the upstream cross-axis identity gap finding)
• lance-graph fix(contract+ogar): mirror OGAR #110 auth domain + bump workspace Cargo.lock — restores q2 build parity #596 (the predecessor D — closed in favor of this)
• sibling D1 (81d08c42) — already mirrored OGAR feat: MoE architecture — 4096 experts, top-128 sparse, 4-group hierarchy 4096 experts → router selects top-128 → each runs 16 internal layers → 4 hierarchical groups where expert outputs compose → collapse → token Architecture maps Qwopus's actual MoE structure: Router: 4096×4096 input distance table (which experts respond?) Expert internals: 256×256 per-layer tables (how does each expert think?) Sparse activation: only top-128 of 4096 fire (97% sparsity) Hierarchical meeting: experts compose at 4 intermediate points Performance: 42s for 3 prompts in release (128×16 MatVec per token) → production needs SIMD batching or GPU Output quality still blocked by attractor collapse ("!" dominates). Needs: thinking style temperature + persona routing + ONNX gate correction. https://claude.ai/code/session_01ChLvBfpJS8dQhHxRD4pYNp #110's auth